Modernizing Health IT: CMS Pledges, AI and the Trust Foundation with Amy Gleason
In this episode of The Dish on Health IT, host Tony Schueth is joined by co-host Alix Goss and special guest Amy Gleason, Strategic Advisor to the Centers for Medicare & Medicaid Services (CMS) and Administrator of the U.S. Department of Government Efficiency (DOGE) Service, for a wide-ranging discussion onhow health IT modernization is evolving under a pledge-driven, incentive-backed federal strategy.
The conversation begins not with policy, but with lived experience.
From Emergency Room to Interoperability Advocate
Amy shares how her early career as an emergency room nurse exposed the dangers of fragmented information. Providers were expected to make critical decisions without access to complete patient histories, while patients, often in pain or distress, were unrealistically asked to recall complex medical details.
That professional frustration became deeply personal when her daughter went more than a year without a diagnosis for a rare autoimmune disease, juvenile dermatomyositis (JDM). Multiple specialists saw pieces of the puzzle, but no one could see the full picture across charts and settings. Amy reflects that if today’s AI tools had been applied to her daughter’s complete longitudinal record, the condition may have surfaced sooner.
That experience shaped her philosophy. Technology must converge with policy and trust in ways that tangibly improve care.
Why Pledges Instead of Rules?
Tony presses on a central theme. Amy has argued that we cannot regulate our way to success. Why pursue voluntary pledges instead of federal rulemaking?
Amy explains her frustration at returning to government in 2025 to find that the interoperability policies she helped draft in 2020 were still not fully effective until 2027. Seven years is an eternity in technology. Meanwhile, the industry had technically complied with numerous mandates, including Meaningful Use, Cures Act APIs, and CMS interoperability rules, yet many workflows still felt broken.
In her view, regulation created a floor but not always real transformation.
The CMS Health Tech Ecosystem Pledge was launched as a different model. The federal government used its convening power to articulate a clear vision and challenge industry to deliver minimum viable products within six to twelve months rather than years.
Initially announced with roughly 60 companies, the pledge initiative has grown to more than 600 participants collaborating in working groups. The three initial patient-focused use cases include:
- Improving data interoperability
- “Killing the clipboard” through digital identity and QR-based sharing
- Leveraging conversational AI and personalized recommendations for chronic conditions such as diabetes and obesity
Amy describes live demonstrations at a Connectathon that showcased OAuth-enabled data retrieval, QR code ingestion into EHR workflows, and AI-powered recommendations built on patient data. The goal is not perfection by the first milestone, but real-world, minimum-viable functionality that can be iteratively improved.
Alix notes that, from the standards community's perspective, this approach aligns with long-standing calls for industry-driven collaboration, though it is still too early to assess its widespread impact.
Carrots, Sticks, and Rural Health
The discussion turns to incentives.
Amy outlines the administration’s carrots and sticks strategy:
- Stick: Enforcement of information blocking, with penalties up to $2 million per occurrence
- Carrots: Financial incentives such as the $50 billion Rural Health Transformation Program and the CMS ACCESS Model, which pays for technology-enabled outcomes
The Rural Health Transformation Program directs funds to states, with the expectation that ecosystem-aligned interoperability and app participation be incorporated into funding proposals. CMS retains oversight and claw-back authority to ensure funds support rural providers.
The ACCESS Model represents a significant shift. Technology-enabled care platforms can register as Medicare Part B providers and be paid for measurable outcomes in tracks such as cardiometabolic disease, musculoskeletal conditions, and behavioral health. Providers remain in the loop and receive compensation for referrals and care plan oversight.
Alix underscores that rural providers face steep financial and workforce constraints. Standards participation, implementation, and technology upgrades require resources that are often scarce. The success of these incentives will depend on whether they reduce the burden rather than add to it.
AI: Evolution, Risk and Reality
AI becomes a central thread of the episode.
Amy compares AI adoption to autonomous vehicle models. Some scenarios allow tightly controlled automation, such as medication refills, while others require a human-in-the-loop for higher-risk decisions. She points to a Utah prescription-refill pilot as an example of bounded automation, where malpractice coverage and clearly defined use cases mitigate risk.
When Tony asks who owns risk in this evolving landscape, Amy emphasizes the need for light but clear regulatory pathways rather than fragmented state-by-state oversight.
Patients, she notes, are already there. Millions are asking health-related questions weekly through AI tools. The more pressing issue is ensuring that those tools are grounded in structured medical data rather than incomplete memory or unverified inputs.
She shares a striking story. Her daughter was excluded from a clinical trial due to a misclassification of ulcerative colitis. By uploading her records into an AI model, they identified a more precise diagnosis, microscopic lymphocytic colitis, which did not disqualify her from the trial. For Amy, this demonstrates both the power and inevitability of AI use.
Alix adds caution. AI is only as strong as the data beneath it. Dirty, inconsistent, and poorly structured data limit performance. Standards and terminology remain essential to fuel high-fidelity models and safeguard trust.
FHIR, Deregulation, and the Data Foundation
The conversation addresses an emerging tension. If regulatory burdens are being reduced, does that signal a less need for structured standards like FHIR?
Amy candidly admits she initially wondered whether AI might reduce the need for FHIR altogether. After discussions with the labs and technologists, she concluded otherwise. Standardized data dramatically improves AI performance and reduces errors.
Deregulation is about removing unnecessary burdens, not abandoning foundational data structures.
Alix reinforces that FHIR enables discrete, normalized data capture that supports both legacy transactions and AI evolution. While future innovations may emerge, today, FHIR remains the backbone of scalable interoperability.
Prior Authorization and HIPAA Modernization
The episode dives into prior authorization modernization across the medical and pharmacy domains.
Amy notes a growing interest among pledge participants in expanding into pharmacy prior authorization testing, diagnostic imaging, real-time benefit checks, and bulk FHIR performance testing.
Alix provides insight into ongoing work within the Designated Standards Maintenance Organizations to incorporate FHIR-based approaches into HIPAA-named standards, particularly for prior authorization. She highlights testing beyond Connectathons, including implementer communities and real-world pilot efforts.
Both stress the importance of public comment periods and industry engagement, describing participation as a civic responsibility for health IT professionals.
Trust as the Core Enabler
The final segment centers on trust.
Amy explains that the ecosystem initiative aims to reinforce trust through:
- Stronger digital identity verification, such as Clear, ID.me, and Login.gov
- Certification frameworks such as CARIN and DIME for patient-facing apps
- A new national provider directory to replace fragmented provider data sources
- Transparency dashboards showing data requests, volumes, and purpose
Rather than replacing frameworks such as TEFCA, she describes the pledge model as an accelerator layered on top of the regulatory floor.
Transparency acts as sunlight, enabling visibility into who is accessing data and for what purpose.
Final Takeaways
In closing, Amy urges providers not to sit on the sidelines. Too often, she says, providers feel change is imposed on them. The pledge environment is designed as an open forum where they can directly shape what works and what does not in real workflows.
Alix echoes the call. Standards require participation. Organizations must allocate budget and staff to engage, comment, and collaborate. It truly takes a village.
Tony concludes by framing the episode’s core message. Regulation establishes baseline expectations, but voluntary movements can demonstrate what is possible before mandates reach the Federal Register.
Across pledges, payment reform, AI evolution, and trust frameworks, the episode underscores a consistent theme. Modernization in health IT depends not only on policy direction but also on shared accountability and active participation from every stakeholder in the ecosystem.
Listeners are reminded that POCP is available to support organizations in understanding the implications of federal initiatives, enforcement priorities, and their strategic implications. Reach out to us to set up an initial consultation.
The episode closes, as always, with the reminder that Health IT is a dish best served hot.
Prefer video? Catch episodes on the POCP YouTube channel
